A single-phase CHECK index operation for db2 entails creating a special sort record for each data record and each index entry, collating the special sort records in a particular way to group together the sort records for each data record and any associated index entries, and performing diagnostic operations on the sorted records.
|
1. A method of testing db2 index integrity, comprising the steps of:
(a) reading, in parallel, (1) a plurality of db2 data records from a db2 table space, and (2) a plurality of db2 index entries from each of at least one db2 indexes associated with said db2 table space, normally each said index entry being nominally associated with exactly one said db2 data record; (b) constructing a sort record for each said data record and each said index entry; (c) collating the sort records for the data records and said sort records for the index entries into a single sequence of sort records to group together (1) the sort records for each data record, with (2) the one or more sort records for said one or more index entries associated with said respective date record; and (d) performing a specified diagnosis routine utilizing said single sequence of sort records as an input.
2. A method of testing db2 index integrity, comprising the steps of:
(a) reading, in parallel, (1) a plurality of db2 data records from a db2 table space, and (2) a plurality of db2 index entries from each of at least one db2 indexes associated with said db2 table space, each said index entry being nominally associated with exactly one said db2 data record; (b) constructing a sort record for each data record, each said sort record respectively including (1) a row identifier field whose value uniquely identifies said respective data record, and (2) an index identifier field whose value indicates that said sort record is associated with a data record; (c) constructing a sort record for each said index entry, each said sort record for said index entries respectively including (1) a field identifying the data record with which said index entry is associated, and (2) a field identifying the db2 index from which the index entry was read; (d) collating the sort records for the data records and said sort records for the index entries into a single sequence of sort records to group together (1) the sort records for each data record, with (2) the one or more sort records for said one or more index entries associated with said respective data record; and (e) performing a specified diagnosis routine utilizing said single sequence of sort records as an input.
3. The method of a specified one of
4. A program storage device encoding a program of instructions, for a general-purpose programmable machine, wherein execution of said instructions by said machine results in performance of the method steps of
5. A program storage device encoding a program of instructions, for a general-purpose programmable machine, wherein execution of said instructions by said machine results in performance of the method steps of
6. A program storage device encoding a program of instructions, for a general-purpose programmable machine, wherein execution of said instructions by said machine results in performance of the method steps of
|
In the prior art, a well-known process that is executed periodically for DB2 installations is to check the indexes of a database. Generally speaking, this entails ensuring that the physical data records (sometimes referred to as rows) are properly indexed. That is, for each data row should have exactly one index entry (e.g., "Last Name" for a personnel data record), and that the key value, in the physical data record or row matches the key value as stored in the appropriate index entry for that data row.
Some nomenclature is introduced here to aid in understanding the terminology used. A principal term used here is that of a "data set." Generally speaking, a data set is a collection of data that is referred to in an operating system environment (e.g., the well-known MVS operating system environment) by a single name, in much the same way as a word processing file might be given a single directory name for easy retrieval of the data in the file (even though the data might physically be stored in a variety of locations on a disk). Examples of MVS data sets include DB2 indexes and DB2 table spaces. Typically, maintenance operations in the DB2 environment involve three steps: First, reading the data to be reorganized (e.g., from a table space or index), often from a variety of different physical locations identified by a data set name. The step of data reading is typically referred to as an UNLOAD process that involves physically copying the data to some other memory or other storage. The second step is that of sorting or otherwise ordering the data to conform to the desired ordering and performing any other desired processing. Finally, the third step is that of rewriting the sorted data to storage (table space or index) designated with the same data set name. The third step is typically referred to as a RELOAD process.
A conventional way of checking DB2 indexes is illustrated in FIGS. 1 and 2. One or more DB2 indexes 100 are UNLOADed in a first processing phase 105, which typically reads the index entries and sorts them into the order of the physical sequence of the rows indexed. A work or intermediate data set 110 is created from the output of the first processing phase 105. In a second processing phase 115, the contents of the work data set 110 are methodically compared with the actual physical data records read from DB2 table space(s) 120 to check the index integrity as described above, and appropriate diagnostic message(s) 125 are displayed. This process is shown in more detail in FIG. 2. The index(es) in question are UNLOADed sequentially at block 200. At block 210, the index(es) are sorted by the data row identifier (sometimes referred to as the row ID or RID). At block 220, the physical data records are read, and at block 230, the respective data records are compared with the corresponding index key values from the index-related work data set 110. At block 240, appropriate conventional diagnostic routines are executed to check the index integrity.
A significant disadvantage of the foregoing prior-art approach to the CHECK INDEX operation is that the various indexes are UNLOADed one after another, and subsequently the physical data records are read. The elapsed time required for the operation thus includes the sum of the elapsed times required for each of the index UNLOAD operations plus the time required for the reading of the associated physical data records. This is especially undesirable in installations which attempt to operate 24 hours a day, seven days a week; it may be highly undesirable to disable a data set, in effect, for the length of time required for a two-phase CHECK INDEX operation. Moreover, such an approach plainly requires additional input/output (I/O) steps, which can have tangible financial costs associated therewith. In some actual systems, storage space may be at a premium and thus the use of storage to create a working or intermediate data set may be undesirable.
A method in accordance with the invention entails performing a single-phase CHECK INDEX operation for a DB2 environment by creating a special SORT record for each data record and each index entry, collating the special SORT records in a particular way to group together the SORT records for each data record and any associated index entries, and performing diagnostic operations on the sorted records.
FIG. 1 is a data flow diagram overview, and
FIG. 2 a flow-chart overview, of a typical prior-art two-phase CHECK INDEX process in a DB2 database.
FIG. 3 is a data flow diagram of an improved single-phase CHECK INDEX process.
FIG. 4 is a flow-chart representation of the same process.
FIG. 5 shows a sort key used in the process.
FIGS. 6 and 7 are pseudocode and flow-chart representations of an illustrative diagnostic routine.
One illustrative embodiment of the invention is described below as it might be implemented by loading of program of instructions (e.g., executable code or compilable or interpretable source code) on a general purpose computer system from a program storage device such as a magnetic tape, a floppy disk, an optical disk, etc., and causing the computer system to execute the program. In the interest of clarity, not all features of an actual implementation are described in this specification. It will of course be appreciated that in the development of any such actual implementation (as in any software development project), numerous implementation-specific decisions must be made to achieve the developers' specific goals and subgoals, such as compliance with system- and business-related constraints, which will vary from one implementation to another. Moreover, it will be appreciated that such a development effort might be complex and time-consuming, but would nevertheless be a routine undertaking of software engineering for those of ordinary skill having the benefit of this disclosure.
Referring to FIGS. 3 and 4, an improved CHECK INDEX operation takes advantage of an innovative use of the system SORT routine (e.g., the standard SORT routine supplied with the IBM MVS operating system or any desired third-party SORT routine). As shown in FIG. 3, the DB2 index(es) 100 and the physical data rows in the DB2 table space 120 are read in parallel and fed to a SORT routine 300. The output of the SORT routine is examined to diagnose potential problems with index integrity.
Referring now to FIG. 4, the method of the invention is shown in more detail. The physical data records to be checked are read from the DB2 table space(s) 120 along with the index records from the DB2 index(es) 100, as shown in block 400. Significantly, these READ operations are performed in parallel, thus reducing the elapsed time required to perform the CHECK INDEX operation.
At block 405, a series of SORT input records is constructed, one for each data row read from a DB2 table space and one for each index entry read from a DB2 index, as illustrated in more detail in FIG. 5. As is well-known to those of ordinary skill, each index entry includes a value identifying the data row that is indexed by the index entry (commonly referred to as the row ID or "RID"). Each SORT input record accordingly includes a row ID: if the SORT input record is for a data row, then the row ID is that of the data row; if the SORT input record is for an index entry, the row ID is that of the data row that is indexed. In the SORT record, the row ID immediately precedes an index identifier for the record in question. Similarly, the table ID (TB ID) identifies the DB2 table associated with the record in question (whether a table space record or an index record). Finally, the key value field of the SORT record is the index key value for the record in question. If the record in question is a physical data record from a DB2 table space, then the value of the index identifier is set at zero; otherwise, the value of the index identifier is a unique non-zero value uniquely identifying the particular DB2 index from which the index entry was read.
Returning to FIG. 4, the SORT records so constructed are fed to the system SORT routine where a two-part sort key is used. The sort key, as shown in FIG. 5, consists of the row ID immediately followed by the index ID. Consequently, the output of the SORT routine is a series of records in row ID order and for each row ID, in index ID order.
As shown in block 415 and FIGS. 6 and 7, the output of the SORT routine is tested, using any convenient diagnostic process, to determine whether any integrity problems exist in the data being tested. FIG. 6 shows a pseudocode representation of one diagnostic process, which is also shown in graphic flow chart form in FIG. 7.
A significant advantage of the method described herein is that the data-record and index-entry data are read in parallel, thus reducing elapsed time. Moreover, because a work data set 110 is not created, I/O resources are thereby conserved.
Hintz, Thomas E., Tenberg, Kerry C.
Patent | Priority | Assignee | Title |
5842197, | Aug 29 1996 | Oracle International Corporation | Selecting a qualified data repository to create an index |
5842208, | Apr 09 1997 | International Business Machines Corporation | High performance recover/build index system by unloading database files in parallel |
5878410, | Sep 13 1996 | Microsoft Technology Licensing, LLC | File system sort order indexes |
5978793, | Apr 18 1997 | International Business Machines Corporation | Processing records from a database |
6065005, | Dec 17 1997 | International Business Machines Corporation | Data sorting |
6081799, | May 21 1998 | International Business Machines Corporation | Executing complex SQL queries using index screening for conjunct or disjunct index operations |
6144970, | Sep 24 1998 | International Business Machines Corporation | Technique for inplace reorganization of a LOB table space |
6272486, | Apr 16 1998 | International Business Machines Corporation | Determining the optimal number of tasks for building a database index |
6282570, | Dec 07 1998 | International Business Machines Corporation | Monitoring a large parallel database through dynamic grouping and sequential sampling |
6295539, | Sep 14 1998 | CA, INC | Dynamic determination of optimal process for enforcing constraints |
6343286, | Sep 24 1998 | International Business Machines Corporation | Efficient technique to defer large object access with intermediate results |
6343293, | Sep 24 1998 | International Business Machines Corporation | Storing the uncompressed data length in a LOB map to speed substring access within a LOB value |
6363389, | Sep 24 1998 | International Business Machines Corporation | Technique for creating a unique quasi-random row identifier |
6366902, | Sep 24 1998 | International Business Machines Corp. | Using an epoch number to optimize access with rowid columns and direct row access |
6427143, | Apr 10 1998 | CA, INC | Method for loading rows into a database table while enforcing constraints |
6470359, | Sep 24 1998 | International Business Machines Corporation | Fast technique for recovering an index on an auxiliary table |
6598041, | Sep 07 2000 | International Business Machines Corporation | Method, system, and program for processing modifications to data in tables in a database system |
6604097, | Sep 07 2000 | International Business Machines Corporation | Method, system, and program for using control data structures when performing operations with respect to a database |
6606617, | Sep 24 1998 | International Business Machines Corporation | Optimized technique for prefetching LOB table space pages |
6643637, | Sep 07 2000 | International Business Machines Corporation | Method, system, and program for using a fetch request to make data available to an application program |
6665678, | Sep 07 2000 | International Business Machines Corporation | Method, system, and program for optimistic concurrency control for scrollable cursors in a database |
6694305, | Sep 07 2000 | International Business Machines Corporation | Method, system, and program for processing modifications to data in tables in a database system |
6694340, | Sep 24 1998 | International Business Machines Corporation | Technique for determining the age of the oldest reading transaction with a database object |
6721731, | Sep 07 2000 | International Business Machines Corporation | Method, system, and program for processing a fetch request for a target row at an absolute position from a first entry in a table |
6754653, | Sep 07 2000 | International Business Machines Corporation | Method, system, and program for processing a fetch request for a target row in a table that precedes as current row |
6775676, | Sep 11 2000 | International Business Machines Corporation; SAP AG | Defer dataset creation to improve system manageability for a database system |
6856996, | Mar 30 2001 | International Business Machines Corporation | Method, system, and program for accessing rows in one or more tables satisfying a search criteria |
6983321, | Jul 10 2000 | BMC SOFTWARE, INC | System and method of enterprise systems and business impact management |
7376665, | Nov 06 2003 | AT&T Intellectual Property I, L P | Systems, methods and computer program products for automating retrieval of data from a DB2 database |
7774458, | Jul 10 2000 | BMC Software, Inc. | System and method of enterprise systems and business impact management |
7933922, | Nov 06 2003 | RAKUTEN GROUP, INC | Systems, methods and computer program products for automating retrieval of data from a DB2 database |
8438173, | Jan 09 2009 | Microsoft Technology Licensing, LLC | Indexing and querying data stores using concatenated terms |
8818953, | Nov 29 2004 | BMC Software, Inc. | Method and apparatus for loading data into multi-table tablespace |
Patent | Priority | Assignee | Title |
4817036, | Mar 15 1985 | Brigham Young University | Computer system and method for data base indexing and information retrieval |
4933848, | Jul 15 1988 | International Business Machines Corporation | Method for enforcing referential constraints in a database management system |
5089985, | Apr 07 1988 | International Business Machines Corporation | System and method for performing a sort operation in a relational database manager to pass results directly to a user without writing to disk |
5121493, | Jan 19 1990 | AMALGAMATED SOFTWARE OF NORTH AMERICA, INC , A CORP OF TEXAS | Data sorting method |
5133068, | Sep 23 1988 | International Business Machines Corporation | Complied objective referential constraints in a relational database having dual chain relationship descriptors linked in data record tables |
5146590, | Jan 13 1989 | International Business Machines Corporation | Method for sorting using approximate key distribution in a distributed system |
5222235, | Feb 01 1990 | BMC SOFTWARE, INC | Databases system for permitting concurrent indexing and reloading of data by early simulating the reload process to determine final locations of the data |
5307484, | Mar 06 1991 | NEW CARCO ACQUISITION LLC; Chrysler Group LLC | Relational data base repository system for managing functional and physical data structures of nodes and links of multiple computer networks |
5404510, | May 21 1992 | Oracle International Corporation | Database index design based upon request importance and the reuse and modification of similar existing indexes |
5497486, | Mar 15 1994 | LOT 19 ACQUISITION FOUNDATION, LLC | Method of merging large databases in parallel |
Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
Dec 16 1993 | BMC Software, Inc. | (assignment on the face of the patent) | / | |||
Feb 16 1994 | HINTZ, THOMAS E | BMC SOFTWARE, INC | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 006927 | /0811 | |
Feb 16 1994 | TENBERG, KERRY C | BMC SOFTWARE, INC | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 006927 | /0811 | |
Sep 10 2013 | BMC SOFTWARE, INC | CREDIT SUISSE AG, CAYMAN ISLANDS BRANCH, AS COLLATERAL AGENT | SECURITY AGREEMENT | 031204 | /0225 | |
Sep 10 2013 | BLADELOGIC, INC | CREDIT SUISSE AG, CAYMAN ISLANDS BRANCH, AS COLLATERAL AGENT | SECURITY AGREEMENT | 031204 | /0225 | |
Oct 02 2018 | Credit Suisse AG, Cayman Islands Branch | BMC ACQUISITION L L C | RELEASE OF PATENTS | 047198 | /0468 | |
Oct 02 2018 | Credit Suisse AG, Cayman Islands Branch | BLADELOGIC, INC | RELEASE OF PATENTS | 047198 | /0468 | |
Oct 02 2018 | BLADELOGIC, INC | CREDIT SUISSE, AG, CAYMAN ISLANDS BRANCH, AS COLLATERAL AGENT | SECURITY INTEREST SEE DOCUMENT FOR DETAILS | 047185 | /0744 | |
Oct 02 2018 | Credit Suisse AG, Cayman Islands Branch | BMC SOFTWARE, INC | RELEASE OF PATENTS | 047198 | /0468 | |
Oct 02 2018 | BMC SOFTWARE, INC | CREDIT SUISSE, AG, CAYMAN ISLANDS BRANCH, AS COLLATERAL AGENT | SECURITY INTEREST SEE DOCUMENT FOR DETAILS | 047185 | /0744 | |
Jun 01 2020 | BLADELOGIC, INC | THE BANK OF NEW YORK MELLON TRUST COMPANY, N A , AS COLLATERAL AGENT | SECURITY INTEREST SEE DOCUMENT FOR DETAILS | 052844 | /0646 | |
Jun 01 2020 | BMC SOFTWARE, INC | THE BANK OF NEW YORK MELLON TRUST COMPANY, N A , AS COLLATERAL AGENT | SECURITY INTEREST SEE DOCUMENT FOR DETAILS | 052844 | /0646 | |
Sep 30 2021 | BLADELOGIC, INC | ALTER DOMUS US LLC | GRANT OF SECOND LIEN SECURITY INTEREST IN PATENT RIGHTS | 057683 | /0582 | |
Sep 30 2021 | BMC SOFTWARE, INC | ALTER DOMUS US LLC | GRANT OF SECOND LIEN SECURITY INTEREST IN PATENT RIGHTS | 057683 | /0582 | |
Jan 31 2024 | ALTER DOMUS US LLC | BMC SOFTWARE, INC | TERMINATION AND RELEASE OF SECURITY INTEREST IN PATENTS | 066567 | /0283 | |
Jan 31 2024 | ALTER DOMUS US LLC | BLADELOGIC, INC | TERMINATION AND RELEASE OF SECURITY INTEREST IN PATENTS | 066567 | /0283 | |
Feb 29 2024 | CREDIT SUISSE AG, CAYMAN ISLANDS BRANCH, AS RESIGNING COLLATERAL AGENT | GOLDMAN SACHS BANK USA, AS SUCCESSOR COLLATERAL AGENT | OMNIBUS ASSIGNMENT OF SECURITY INTERESTS IN PATENT COLLATERAL | 066729 | /0889 |
Date | Maintenance Fee Events |
May 15 2000 | M183: Payment of Maintenance Fee, 4th Year, Large Entity. |
Apr 20 2004 | M1552: Payment of Maintenance Fee, 8th Year, Large Entity. |
May 16 2008 | M1553: Payment of Maintenance Fee, 12th Year, Large Entity. |
Date | Maintenance Schedule |
Nov 26 1999 | 4 years fee payment window open |
May 26 2000 | 6 months grace period start (w surcharge) |
Nov 26 2000 | patent expiry (for year 4) |
Nov 26 2002 | 2 years to revive unintentionally abandoned end. (for year 4) |
Nov 26 2003 | 8 years fee payment window open |
May 26 2004 | 6 months grace period start (w surcharge) |
Nov 26 2004 | patent expiry (for year 8) |
Nov 26 2006 | 2 years to revive unintentionally abandoned end. (for year 8) |
Nov 26 2007 | 12 years fee payment window open |
May 26 2008 | 6 months grace period start (w surcharge) |
Nov 26 2008 | patent expiry (for year 12) |
Nov 26 2010 | 2 years to revive unintentionally abandoned end. (for year 12) |